ElevenLabs released version 4 of its speech model, adding more expression control and support for 90 languages.

What it means

ElevenLabs has released version 4 of its speech synthesis model, along with a faster 'v4 Turbo' variant. The new models offer finer control over vocal expression (e.g., tone, emotion, pacing) and support over 90 languages, built on a new architecture for improved quality and faster voice cloning.

Accuracy & sourcesGrounded

The claim is well-sourced from ElevenLabs' official announcement, reported by TechCrunch and corroborated by multiple outlets. The details about v4 and v4 Turbo models, expression control, and 90+ languages are consistent across sources.

How new

Genuinely new: published on September 28, 2026, covering a launch that same day.

Go deeper